Test cases provided by Anton Lokhmot.

llvm-svn: 135322

Needed to avoid compile error after the patch "Convert CallInst and
InvokeInst APIs to use ArrayRef.

Contributed by: Sebastian Pop <sebpop@gmail.com>

llvm-svn: 135321

llvm-svn: 135320

Get rid of the separate opcodes for the Darwin versions of tBL, tBLXi, and tBLXr, using pseudo-instructions to lower to the single final opcode.  Update the ARM disassembler for this change.

llvm-svn: 135319

and just use the ones from TargetLowering directly.

llvm-svn: 135318

llvm-svn: 135317

llvm-svn: 135316

o GetChildAtIndex, and
o GetValueForExpressionPath

llvm-svn: 135315

llvm-svn: 135314

1) Make non-legal 256-bit loads to be promoted to v4i64. This lets us
canonize the loads and handle things the same way we use to handle
for 128-bit registers. Despite of what one of the removed comments
explained, the load promotion would not mess with VPERM, it's only a
matter of doing the appropriate bitcasts when this instructions comes
to be introduced. Also make LOAD v8i32 legal.

2) Doing 1) exposed two bugs:
- v4i64 was being promoted to itself for several opcodes (introduced
in r124447 by David Greene) causing endless recursion and the stack to
explode.
- there was no support for allOnes BUILD_VECTORs and ANDNP would fail to
match because it was generating early target constant pools during
lowering.

3) The testcases are already checked-in, doing 1) exposed the
bugs in the current testcases.

4) Tidy up code to be more clear and explicit about AVX.

llvm-svn: 135313

comming together with other tests.

llvm-svn: 135312

llvm-svn: 135311

Add tests for CFRefReport's path notes, and fix a few typos and non-standard terminology ('+0 retain counts') caught by the tests.

llvm-svn: 135310

llvm-svn: 135309

conflicts with a to be produced temp filename.
rdar://9724657

llvm-svn: 135308

This gets rid of some of the gory splitting details in RAGreedy and
makes them available to future SplitKit clients.

Slightly generalize the functionality to support multi-way splitting.
Specifically, SplitEditor::splitLiveThroughBlock() supports switching
between different register intervals in a block.

llvm-svn: 135307

to allow clients to specify that they've already (correctly) loaded
declarations, and that no further action is needed. 

Also, make sure that we clear the "has external lexical declarations"
bit before calling FindExternalLexicalDecls(), to avoid infinite
recursion.

llvm-svn: 135306

llvm-svn: 135305

llvm-svn: 135304

llvm-svn: 135303

llvm-svn: 135302

id x = ...
@try {
 ...
} @finally {
 [x release];
}

Migrator will drop the release. It's better to change it to "x = 0" in a @finally to avoid leak when exception is thrown.

rdar://9398256

llvm-svn: 135301

llvm-svn: 135300

llvm-svn: 135299

llvm-svn: 135298

not isEquality().

llvm-svn: 135296

Add some more docstrings for SBCompileUnit and SBBreakpoint, plus incorporate the doxgen doc block of

SBValue::GetChildAtIndex(uint32_t idx, 
                         lldb::DynamicValueType use_dynamic,
                         bool can_create_synthetic);

into the SBValue docstrings.

llvm-svn: 135295

llvm-svn: 135294

llvm-svn: 135293

represent pointers and arrays by adding an extra parameter to the 

SBValue
SBValue::GetChildAtIndex (uint32_t idx, 
                         DynamicValueType use_dynamic, 
                         bool can_create_synthetic);

The new "can_create_synthetic" will allow you to create child values that
aren't actually a part of the original type. So if you code like:

int *foo_ptr = ...

And you have a SBValue that contains the value for "foo_ptr":

SBValue foo_value = ...

You can now get the "foo_ptr[12]" item by doing this:

v = foo_value.GetChiltAtIndex (12, lldb.eNoDynamicValues, True);

Normall the "foo_value" would only have one child value (an integer), but
we can create "synthetic" child values by treating the pointer as an array.

Likewise if you have code like:

int array[2];

array_value = ....

v = array_value.GetChiltAtIndex (0);     // Success, v will be valid
v = array_value.GetChiltAtIndex (1);     // Success, v will be valid
v = array_value.GetChiltAtIndex (2);     // Fail, v won't be valid, "2" is not a valid zero based index in "array"

But if you use the ability to create synthetic children:

v = array_value.GetChiltAtIndex (0, lldb.eNoDynamicValues, True);     // Success, v will be valid
v = array_value.GetChiltAtIndex (1, lldb.eNoDynamicValues, True);     // Success, v will be valid
v = array_value.GetChiltAtIndex (2, lldb.eNoDynamicValues, True);     // Success, v will be valid

llvm-svn: 135292

llvm-svn: 135291

Remove VMOVDneon and VMOVQ, which are just aliases for VORR.  This continues to simplify the path towards an auto-generated disassembler.

llvm-svn: 135290

llvm-svn: 135289

- No ELF or COFF implementation yet, I don't have a way to test that.
  Should be straightforward to add though.

llvm-svn: 135288

llvm-svn: 135287

is named after a common idiom (i.e., memset/memcpy).  Otherwise, we can run into 
infinite recursion.  Ideally, the user should use the correct -fno-builtin flag,
but in case they don't we should play nicely.
rdar://9763412

llvm-svn: 135286

llvm-svn: 135285

llvm-svn: 135284

Remove unnecessary duplicate instruction definitions that simply overloaded the type of VEXT.  This can be achieved with a Pat definition, and is much more disassembler friendly.

llvm-svn: 135283

llvm-svn: 135282